Machine Learning Techniques—Reductions Between Prediction Quality Metrics

  • Alina Beygelzimer
  • John Langford
  • Bianca Zadrozny

Machine learning involves optimizing a loss function on unlabeled data points given examples of labeled data points, where the loss function measures the performance of a learning algorithm. We give an overview of techniques, called reductions, for converting a problem of minimizing one loss function into a problem of minimizing another, simpler loss function. This tutorial discusses how to create robust reductions that perform well in practice. The reductions discussed here can be used to solve any supervised learning problem with a standard binary classification or regression algorithm available in any machine learning toolkit. We also discuss common design flaws in folklore reductions.


Loss Function Fault Diagnosis Quantile Regression Machine Learn Technique Neural Information Processing System 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    N. Ailon and M. Mohri (2007) An Efficient Reduction of Ranking to Classification, New York University Technical Report, TR-2007-903.Google Scholar
  2. 2.
    E. Allwein, R. Schapire, and Y. Singer (2000) Reducing multiclass to binary: A unifying approach for margin classifiers, Journal of Machine Learning Research, 1:113–141.CrossRefMathSciNetGoogle Scholar
  3. 3.
    A. Asuncion, D. Newman (2007) UCI Machine Learning Repository,, University of California, Irvine.Google Scholar
  4. 4.
    N. Balcan, N. Bansal, A. Beygelzimer, D. Coppersmith, J. Langford, and G. Sorkin (2007) Robust reductions from ranking to classification, Proceedings of the 20th Annual Conference on Learning Theory (COLT), Lecture Notes in Computer Science 4539: 604–619.CrossRefMathSciNetGoogle Scholar
  5. 5.
    A. Beygelzimer, V. Dani, T. Hayes, J. Langford, and B. Zadrozny (2005) Error limiting reductions between classification tasks, Proceedings of the 22nd International Conference on Machine Learning (ICML), 49–56.Google Scholar
  6. 6.
    A. Beygelzimer, J. Langford, and P. Ravikumar (2008) Filter trees for cost sensitive multiclass classification.Google Scholar
  7. 7.
    A. Beygelzimer, J. Langford, and P. Ravikumar (2008) Error Correcting Tournaments.Google Scholar
  8. 8.
    A. Beygelzimer, J. Langford, B. Zadrozny (2005) Weighted One-Against-All, Proceedings of the 20th National Conference on Artificial Intelligence (AAAI), 720–725.Google Scholar
  9. 9.
    E. Bredensteiner and K. Bennett (1999) Multicategory classification by Support Vector Machines, Computational Optimization and Applications, 12(1-3): 53–79.MATHCrossRefMathSciNetGoogle Scholar
  10. 10.
    L. Breiman (1996) Bagging predictors, Machine Learning, 26(2):123–140.Google Scholar
  11. 11.
    K. Crammer and Y. Singer (2001) On the algorithmic implementation of multiclass mernel-based vector machines, Journal of Machine Learning Research 2: 265–292.CrossRefGoogle Scholar
  12. 12.
    K. Crammer and Y. Singer (2002) On the learnability and design of output codes for multiclass problems, Machine Learning, 47, 2-3: 201–233.MATHCrossRefGoogle Scholar
  13. 13.
    T. Dietterich and G. Bakiri (1995) Solving multiclass learning problems via error-correcting output codes, Journal of Artificial Intelligence Research, 2: 263–286.MATHGoogle Scholar
  14. 14.
    P. Domingos (1999) MetaCost: A general method for making classifiers cost-sensitive, Proceedings of the 5th International Conference on Knowledge Discovery and Data Mining (KDD), 155–164.Google Scholar
  15. 15.
    C. Drummond and R. Holte (2000) Exploiting the cost (in)sensitivity of decision tree splitting criteria, Proceedings of the 17th International Conference on Machine Learning (ICML), 239–246.Google Scholar
  16. 16.
    C. Elkan (2001) The foundations of cost-sensitive learning, Proceedings of the 17th International Joint Conference on Artificial Intelligence (IJCAI), 973–978.Google Scholar
  17. 17.
    Y. Freund, Y. Mansour and R. Schapire (2004) Generalization bounds for averaged classifiers, The Annals of Statistics, 32(4): 1698–1722.MATHMathSciNetGoogle Scholar
  18. 18.
    Y. Freund and R. Schapire (1997) A decision-theoretic generalization of on-line learning and an application to boosting, Journal of Computer and System Sciences, 55(1): 119–139.MATHCrossRefMathSciNetGoogle Scholar
  19. 19.
    Y. Guermeur, A. Elisseeff, and H. Paugam-Moisy (2000) A new multi-class SVM based on a uniform convergence result, Proceedings of the IEEE International Joint Conference on Neural Networks 4, 183–188.Google Scholar
  20. 20.
    V. Guruswami and A. Sahai (1999) Multiclass learning, boosting, and error-correcting codes, Proceedings of the 12th Annual Conference on Computational Learning Theory (COLT), 145–155.Google Scholar
  21. 21.
    T. Hastie and R. Tibshirani (1998) Classification by pairwise coupling, Advances in Neural Information Processing Systems (NIPS), 507–513.Google Scholar
  22. 22.
    L. Huang, X. Nguyen, M. Garofalakis, J. Hellerstein, M. Jordan, A. Joseph, and N. Taft (2007) Communication-efficient online detection of network-wide anomalies, Proceedings of the 26th Annual IEEE Conference on Computer Communications (INFOCOM), 134–142.Google Scholar
  23. 23.
    A. Kalai and R. Servedio (2003) Boosting in the presence of noise, Proceedings of the 35th Annual ACM Symposium on the Theory of Computing (STOC), 195–205.Google Scholar
  24. 24.
    M. Kearns (1998) Efficient noise-tolerant learning from statistical queries, Journal of the ACM, 45:6, 983–1006.MATHCrossRefMathSciNetGoogle Scholar
  25. 25.
    R. Koenker and K. Hallock (2001) Quantile regression, Journal of Economic Perspectives, 15, 143–156.CrossRefGoogle Scholar
  26. 26.
    J. Langford and A. Beygelzimer (2005) Sensitive Error Correcting Output Codes, Proceedings of the 18th Annual Conference on Learning Theory (COLT), 158–172.Google Scholar
  27. 27.
    J. Langford, R. Oliveira and B. Zadrozny (2006) Predicting conditional quantiles via reduction to classification, Proceedings of the 22nd Conference in Uncertainty in Artificial Intelligence (UAI).Google Scholar
  28. 28.
    J. Langford and B. Zadrozny (2005) Estimating class membership probabilities using classifier learners, Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics.Google Scholar
  29. 29.
    J. Langford and B. Zadrozny (2005) Relating reinforcement learning performance to classification performance, Proceedings of the 22nd International Conference on Machine Learning (ICML), 473–480.Google Scholar
  30. 30.
    Y. Lee, Y. Lin, and G. Wahba (2004) Multicategory support vector machines, theory, and application to the classification of microarray data and satellite radiance data, Journal of American Statistical Association, 99: 67–81.MATHCrossRefMathSciNetGoogle Scholar
  31. 31.
    C. Hsu and C. Lin (2002) A comparison of methods for multi-class support vector machines, IEEE Transactions on Neural Networks, 13, 415–425.CrossRefGoogle Scholar
  32. 32.
    D. Margineantu (2002) Class probability estimation and cost-sensitive classification decisions, Proceedings of the 13th European Conference on Machine Learning, 270–281.Google Scholar
  33. 33.
    M. Mesnier, M. Wachs, R. Sambasivan, A. Zheng, and G. Ganger (2007) Modeling the relative fitness of storage, International Conference on Measuremen and Modeling of Computer Systems (SIGMETRICS), 37–48.Google Scholar
  34. 34.
    J. von Neumann (1951) Various techniques used in connection with random digits, National Bureau of Standards, Applied Mathematics Series, 12: 36–38.MathSciNetGoogle Scholar
  35. 35.
    J. Platt (1999) Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods. In A. Smola, P. Bartlett, B. Schölkopf, and D. Schuurmans, editors, Advances in Large Margin Classifiers, 61–74.Google Scholar
  36. 36.
    J. Platt, N. Cristiani and J. Shawe-Taylor (2000) Large margin DAGs for multiclass classification, Advances of Neural Information Processing Systems, 12: 547–553.Google Scholar
  37. 37.
    J. Platt, E. Kiciman and D. Maltz (2008) Fast variational inference for large-scale internet diagnosis, Advances in Neural Information Processing Systems 20.Google Scholar
  38. 38.
    R. Rifkin and A. Klautau (2004) In defense of one-vs-all classification, Journal of Machine Learning Research, 5: 101–141.MathSciNetGoogle Scholar
  39. 39.
    I. Rish, M. Brodie and S. Ma (2002) Accuracy versus efficiency in probabilistic diagnosis, Proceedings of National Conference on Artificial Intelligence (AAAI), 560–566.Google Scholar
  40. 40.
    I. Takeuchi, Q. Le, T. Sears, and A. Smola (2006) Nonparametric quantile estimation, The Journal of Machine Learning Research, 7, 1231–1264.MathSciNetGoogle Scholar
  41. 41.
    G. Tesauro, R. Das, H. Chan, J. Kephart, D. Levine, F. Rawson, and C. Lefurgy (2008) Managing power consumption and performance of computing systems using reinforcement learning, Advances in Neural Information Processing Systems 20.Google Scholar
  42. 42.
    V. Vapnik (1998) Statistical Learning Theory, John Wiley and Sons.Google Scholar
  43. 43.
    H. Wang, J. Platt, Y. Chen, R. Zhang, and Y. Wang (2004) Automatic Misconfiguration Troubleshooting with PeerPressure, Proceedings of the 6th Symposium on Operating Systems Design and Implementation, (2004). Also in Proceedings of the International Conference on Measurements and Modeling of Computer Systems, SIGMETRICS 2004, 398–399.Google Scholar
  44. 44.
    J. Weston and C. Watkins (1998) Multiclass support vector machines, Proceedings of the 11th European Symposium on Artificial Neural Networks, 219–224.Google Scholar
  45. 45.
    I. Witten and E. Frank (2000) Data Mining: Practical machine learning tools with Java implementations, Morgan Kaufmann,
  46. 46.
    B. Zadrozny and C. Elkan (2001) Learning and making decisions when costs and probabilities are both unknown, Proceedings of the 7th International Conference on Knowledge Discovery and Data Mining (KDD), 203–213.Google Scholar
  47. 47.
    B. Zadrozny, J. Langford and N. Abe (2003) Cost-sensitive learning by cost-proportionate example weighting, Proceedings of the 3rd IEEE International Conference on Data Mining (ICDM), 435–442.Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2008

Authors and Affiliations

  • Alina Beygelzimer
    • 1
  • John Langford
    • 2
  • Bianca Zadrozny
    • 3
  1. 1.IBM Thomas J. Watson Research CenterHawthorne
  2. 2.Yahoo! ResearchNew York
  3. 3.Fluminense Federal UniversityBrazil

Personalised recommendations